Académique Documents
Professionnel Documents
Culture Documents
compiled by
Dr. Aly N. El-Bahrawy
From
Chapter 1
Civil and Environmental Systems Engineering
C. ReVelle, E. Whitlatch, J. Wright
Pearson Prentice Hall, 2004
Introduction
• Everyone’s lifetime is filled with an endless
string of opportunities and challenges to
decide or choose among alternative courses of
action
• Engineers should present alternatives that
serve the interests of their clients
• There are powerful tools that enhance the
decision-making process
Introduction
• Mathematical programming and linear
programming have wide application in
engineering
• The course offer application in large quantities
to motivate the student to learn the
methodologies
• Simple computer programs are used to solve
the problems of optimization
Explaining Systems Analysis
• Mathematical models are built using
differential and difference equations to
explain natural phenomena
• Such models are called collectively ‘descriptive
models’
Explaining Systems Analysis
• After World War II, mathematics for decision
making and invention of digital computers
have enhanced the decision making process
• Models that use such decision making
mathematics are called ‘prescriptive models’
Explaining Systems Analysis
• The question for descriptive models is: ‘If I
follow this course of action, what will
happen?’
• The question for prescriptive models is: ‘What
is the best course of action that I might
follow?’
Explaining Systems Analysis
• Models can be classified as deterministic or
stochastic
• Deterministic models (the number of table
tops of a given size that can be cut from 4 by 8
foot sheet of plywood) have a single outcome
• Stochastic models (maximum temperature
during the month of August) have multiple
outcomes
Deterministic Optimization …
• Linear Programming
– Linear objective and linear constraints
• Quadratic Programming
– Quadratic objective function
• Gradient Methods
– Follow slopes of objective function
• Optimal Control Theory
– Optimal trajectory to achieve some goal
… Deterministic Optimization
• Dynamic Programming
– Problems with a number of time stages
• Multiobjective Programming
– Derives tradeoffs between objectives
• Integer Programming
– Only integer-valued decisions
• Binary Programming
– Only 0 or 1 decisons
Mathematical Programming
• The term means optimization methods and its
subspecialties (linear, dynamic, etc.)
• The programming in the term means
scheduling, the setting of an agenda, or the
creation of a plan of activities.
• Confusion exists with computer programinng
since all optimization requires use of digital
computation.
A History of Systems and Optimization
• Subject to
• and
Definitions
• The program has n variables and m
constraints, j is index for decision, and i index
for resources
• cj is profit (for activity j)
• aij is coefficient (resource i consumed for
activity j)
• bi is amount of resource i available
Decision Model Settings
• We need to collect data needed
• We need to articulate objective and
constraints on acceptable performance and
decision variables
Problem 1 Statement
• Siting of a system of health clinics in a rural
region of a developing country
– 10 health clinics
– All communities over 1000 reach a clinic within
less than 10 miles of travel
– Clinic sited in a community 10,000 or more
• Data needed
– Population of communities, those with 10,000 at
least and distance between them.
Problem 1 Articulation
• Objective
– Average distance travelled as small as possible
• Constraints
– Clinics at communities with 10,000 at least
– At least one site within 10 miles have health clinic
– Only 10 clinics can be sited
• Decision variables
– Which communities assigned to which clinics and
which communities house the clinics
Problem 2 Statement
• Building a reservoir to supply water to a city
• Data needed
– Volumetric demand for water in each month
– For each month, the worst historical sequence of
inflows to the reservoir
– Long records of historical streamflows
Problem 2 Articulation
• Objective
– Smallest capacity reservoir to provide projected
demand for worst drought recorded
• Decision variables
– Reservoir capacity and storages at the end of each
month
• Constraints
– Water stored less than reservoir capacity
– Water in excess of capacity released to stream
– End of month content is greater than zero
Problem 3 Statement
• Upgrading the links of highway network in a
developing nation
• Data needed
– Cost of link built or upgraded
– Population of cities connected to network
– Projected trade between cities and export from
the system
Problem 3 Articulation
• Objective
– Maximize the population connected directly by
ready access
– Maximize the potential trade between cities
• Decision variables
– Which cities to connect or include in the network
– Which links to improve in planned upgrade
• Constraints
– Specified budget from Transport Ministry
Problem 4 Statement
• Positioning of fire stations in an urban area
• Data needed
– Partitioned sectors of the urban area
– Travel time between all centroids
Problem 4 Articulation
• Objective
– The least number of fire stations to achieve 5 min
performance
• Design variable
– 1 or 0 to site or not a fire station in a given sector
• Constraints
– Each sector has fire station 5 min travel time from
centroid
Problem 5 Statement
• Expansion of an electric power system for a
province
• Data needed
– Projection for demand of electric power for each
city
– Cost to build and operate hydroelectric, coal-fired
and nuclear power plants
– Power losses along the segments of the network
Problem 5 Articulation
• Objective
– Least cost (building and operation) of power
plants to meet demands for power
• Decision variables
– 1 or 0 denoting to establish or not a plant type k
at site i built to size j
• Constraints
– Each city assigned sufficient power from among all
plants
Course Project
• Choose a topic from the course contents
• Discuss your choice with me
• Submit a report and make a presentation
• You can choose an interesting Linear
Programming or other mathematical models
application